STCLOUD_cloud computing continuation - cloud computing defined, infra as sw, trad comp model, cloud comp model, levels of virtualization, deployment models

original file copied

Cloud Computing Defined

Cloud computing is the on-demand delivery of computing/IT resources via the network (over the cloud or local network)

• the original was AWS's definition, they are a public cloud service → what about private cloud? the original definition cannot
• cloud computing is a usage of virtualization, a methodology
• cloud computing can be public or private, or hybrid
• private cloud can just be local network
• pay-as-you-go is incorrect if you are on a private cloud setup
• network is our I/O instead of physical devices

Infrastructure as software

• cloud computing (because of virtualization) enables you to stop thinking of your infrastructure as hardware, and instead think of and use it as software

• rent, renovation, physical equipment, issue with physical space
  when virtualized...
• flexibility, scalability, recon figure, backup, portable, resource maximization, isolation

Traditional Computing Model

• infrastructure as hardware
• hardware solutions
  • require space, staff, physical security, planning, capital expenditure
    • capital expenditure is an initial investment or upfront costs
  • have a long hardware procurement cycle (canvas, supplies, bids, signatory, release, shipping, setup, config, usage)
    • when you purchase something, there's a delay in between wanting it and having it
  • require you to provision capacity by guessing theoretical maximum peaks

Under- and over-utilization

Cloud Computing Model → Uses virtualization

• infrastructure as software → you pay a premium and gain certain benefits
• software solutions:
  • are flexible (agility, scale)
  • can change more quickly, easily, and cost-effectively than hardware solutions
    • you don't have to setup the servers, rack and stack them, and other time-consuming setup for physical equipment
  • eliminate the undifferentiated heavy-lifting tasks

Cloud Service Models → Levels of virtualization

• virtualization has different types
• virtualization is a form of abstraction and translation
• the level of abstraction may be different (containers are OS virtualizaiton, VMs are hardware-level) based on the level of virtualization
• so cloud computing, we want to split the services into different categories and levels of virtualization

• IaaS is a service model, allows us to perceive computing resources
  • closest to traditional data centers
  • base computing resources: traditional IT resources
  • virtual servers, VMs, containers, storage, networks, firewalls
  • CPU cores, memory, storage
  • those who will use it will need the most amount of flexibility and control: IT, operations, network security
• PaaS
  • for control panels/cpanels → lets you choose applications & respective versions
  • applications give you flexibility, the version is for compatibility
  • you can choose the programming language (python, node, JS, JQuery, etc., Apache webserver, HTML, nginx, mysql) SCP Secure Copy Protocol
  • those who will use it are devs, people in devops
• SaaS
  • you're a business person or generic user who wants a website or application → user or subscriber of a particular service
  • you don't get to choose the version or webserver you're gonna use
  • ex. wordpress, wix, gdocs, gdrive
  • register to use (accounts)
  • user configuration, user preferences

Cloud Computing Deployment Models

where is it being deployed, how is it being used?

• Cloud (Public cloud) - mostly accessible via the internet, you subscribe/rent resources, global scale (ex. services: Azure, AWS, Google Cloud, Alibaba, Oracle, etc.)
  • concept of trust (you have to trust the cloud provider you choose, consider the security implications)
• Private Cloud/On-premises - cloud service that is deployed on-premises (ex. CCS Cloud), usually owned by a single organization/single tenant, accessible via intranet or internet, high capex but low opex, more security because it's just local (you can just cut off attackers by cutting the internet access)
• Hybrid Cloud - public cloud and private cloud working together; bale just having a combination of both
  • 2 servers in private cloud → high capex low opex → this is for your minimum capacity
  • for maximum peaks, rent 1-4 servers on the public cloud. if you don't need them, don't rent them
  • more flexibility while getting the certain advantages that only either public has or private has
  • recall that a fully cloud setup comes out expensive

In terms of security... there's the Shared Responsibility Model

• the vendor → focused on the infra side of cloud
• the user → focused on the configuration (if you're using SaaS with OS vulnerability, the fault is on the cloud provider) (If you're using SaaS and there's an issue with user password → the fault is on the user)
  • if its something you can't configure and the cloud handles it, it's the fault of the provider
  • choosing SaaS means you will have less responsibility, if you chose IaaS then you have more responsibility on the password
• if you understand the implications of security, then you might want to go for a private cloud
• private cloud gives you more control but you will have more responsibility on more aspects (like security)
• in public and hybrid cloud, if the vendor gets compromised you also get compromised.
• it's up to your understanding and skillset, and level of responsibility that you want

next

• similarity of data centers, aws module 2